tool-evaluation-and-adoption

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is designed as a defensive utility that prioritizes user safety during software adoption. It explicitly forbids installation without authorization and includes specific precautions against piping remote scripts directly to a shell.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves ingesting and analyzing untrusted third-party data from external repositories and documentation.
  • Ingestion points: Untrusted data enters the context in Step 1 (SKILL.md) and during Reconnaissance (references/reconnaissance.md) when the agent clones repositories and reads README files or documentation sites.
  • Boundary markers: The skill utilizes a structured procedural framework that separates analysis from execution, requiring the agent to wait for a decision after presenting findings.
  • Capability inventory: The skill uses git clone, git worktree (in references/adoption-verification.md), and potentially package managers or shell-based installation commands in Step 7.
  • Sanitization: There are explicit instructions to manually inspect installation scripts and verify checksums for release assets before proceeding with any installation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 04:19 AM
Security Audit — agent-trust-hub — tool-evaluation-and-adoption